基于相干系数的低相干区域识别

Low- coherence Area Identification Based on the Coherence Coefficient

由于成像机理的影响，SAR影像不可避免地存在低相干区域，如道路、湖泊以及山的背坡等。低相干区域回波信号较弱或者没有回波信号、数据相干性差、无法用于后续测绘产品生成，必须将其识别出来。本文在传统滤波与闽值分割识别方法的基础上，引入均值、方差作为判据，并辅以误判区域的剔除策略，有效完成了InSAR数据中低相干区域的识别；同时，通过对机载InSAR数据处理实验，验证了该算法的有效性。

Due to the imaging mechanism, the existence of low - coherence areas such as road, lake and back slope of mountain is inevitable in SAR image. It is necessary to identify the low - coherence areas because the intensity of echo in those areas is very low or even vanished, thus the data is incoherent and useless for subsequent generating of surveying and mapping products. Based on the traditional filtering and segmenting method, the paper presents a new method that can effectively identify the low - coherence area by using mean and variance as criterion and deleting the misjudged area. Besides the paper testifies that the method is effective by the experiments using airborne InSAR data.